In later generations, many people who were not interested in science, engineering and engineering technology felt that artificial intelligence was just a tool and an application.

Even if AI has the potential to take away your job, everyone will be afraid, or curse, and what to do after cursing, and the matter will be over.

However, Gu Wan knows that there is an iron law of artificial intelligence that has been regretted by countless people in later generations for not knowing earlier: this iron law is easy to understand, as long as you don't have prejudices and resistance to technical things, then even if you are a pure liberal arts student, you can easily understand this iron law and benefit from it.

To put it mildly, even if you think of it as chicken soup, as long as you have a system in your head, the system gives you a task to tell you that you can only drink one bowl of chicken soup produced by the entire human race in the 21st century, and the rest must be poured out, and if you drink one more bowl, the system will wipe you out.

So, what do you choose at that moment?

You should throw out all the other chicken broth produced by all mankind and drink only this one bowl.

This is a bowl of scientific chicken soup that has been scientifically rigorously proven and can benefit people's learning ability for a lifetime.

The tonic it contains belongs to the grade of Chaowen Dao and Xi Ke Death, and the entrance is the internal force of Yijiazi. In martial arts, the protagonist is not worthy of understanding this way of the origin of the universe without jumping off the cliff ten times.

And Gu Yu is at this moment, he is describing how to measure this bowl of scientific chicken soup to Ma Yiyi.

It's just that many prophetic words need to be embellished before they can be spoken, so it becomes the following kind of inducing dialogue:

"Have you ever thought about a question: you have been studying for thirteen years in your life, and your learning efficiency has always been so high and stable?

Isn't there a time when the study efficiency is low, the self-study night has not made any progress, and the paper has not been fruitful?"

Ma Yiyi immediately felt related.

Although Xueba and Xueshen have stronger learning ability than normal people, they are also more sensitive to changes in efficiency. Sometimes after learning for a while and getting nothing to gain, you will be more anxious than ordinary people, and then adjust your learning methods and pace.

If he is a heartless scumbag, he may do repetitive work that has no harvest all night, and he is happy not to think there is any problem.

"Of course, I am often dissatisfied with my learning efficiency, and sometimes I feel that the teacher is wasting my time in class. Ma Yiyi said empathetically.

Gu joked: "This question, in the final analysis, is because the pace of learning does not accurately match your ability area." You've also studied psychology and cognitive neuroscience for a semester, and you should be familiar with the following concepts, namely the 'learning zone', 'comfort zone' and 'panic zone' in learning. ”

Ma Yiyi: "Of course I know this, and the comfort zone is something I fully understand." For example, we are doing papers a lot of times before the college entrance examination, and some simple questions are already ripe, even if it is to deepen the impression, it is enough to practice three or four times a month.

But when I was talking about naval battles and doing mock papers, I had to practice dozens or even hundreds of times a month, let alone three or five times, and I was very impatient when I got emotionally hairy later. This is the comfort zone, and I haven't improved since I practiced, so I understand everything.

As for the panic zone, it's just a paper to read, and some topics don't understand anything, and I don't have a clue at all. Not only do I not know how to solve it, but I don't even know the prior knowledge required to solve it, and I completely listen to the book of heaven.

And the final learning zone, which is somewhere between the comfort zone and the panic zone, is just right for you. It's a little challenging, there are some things you don't understand, but as long as you work hard, with the help of the existing knowledge in your existing knowledge structure system, you can graft, induct, deduce, and deduce, and you can solve this problem.

When you are in the study area, if the proportion of people who don't understand is too high, you will panic and get tired of learning. If the proportion of people who don't understand is too low, they will be tired and too lazy to distract. Only when the proportion of people who do not understand is just right, it is easiest to enter the most efficient learning state, according to the theory of the great psychologist Mihaly, this state is called 'flow'. ”

"The foundation of psychology and cognitive neuroscience is not bad. Gu Play praised his girlfriend,

Indeed, as a newcomer who has only been studying psychology for a semester, it is already very awesome to have this kind of opinion.

Then Gu Yu changed his words: "However, in the past, we all thought that flow was a state that could be encountered but not sought, but soon, with the analysis of machine learning, we will find that there is a scientific optimal solution to enter the flow or the most efficient learning state."

This optimal solution requires us to adjust the ratio of right and wrong, and the ratio of difficulty and difficulty of a volume, a learning, and a machine big data training set to an optimal mystical value. Only when this value is tailored, both humans and machines can achieve the most perfect learning efficiency. Even if you are a scumbag, you can feel the efficiency of forgetting the eyes of the sky in your studies. ”

"How do you do that?" Ma Yiyi had forgotten that she was talking about academic issues.

……

yes, how?

On Earth, in 2018, two experts in artificial intelligence algorithms from the University of Arizona and Brown University gave the optimal solution.

Their conclusion is that when the ratio of right and wrong is controlled in a machine learning training set at %, the machine learning algorithm can achieve the highest efficiency and the fastest progress.

For example, let the computer use artificial intelligence image recognition to identify 10,000 pictures that resemble cats, and learn "how to tell if the thing in the picture is a cat".

At this time, you have to take 8413 pictures of real cats, and 1587 pictures of cats and non-cats, to feed data to artificial intelligence, then after the machine learns these 100000 pictures and gets the right and wrong score, the improvement is the most.

This is the optimal solution of the laws of natural mathematics, in other words, if you take 8414 real cat pictures and 1586 cat-like non-cat pictures to feed, after the machine eats these 10,000 big data, the amount of progress will be slightly lower than that of the previous group.

How did this data come from? Human algorithmicists, since 2010, when Google began to practice deep learning, have been experimenting with all human algorithm experts for eight years.

What's even more amazing is that the earthlings later did more in-depth experiments and found that this learning efficiency mechanism is really not only applicable to machine learning, but also to the human brain.

In terms of "trial-and-error learning", or early cognition, the human brain is the same as deep learning, which simulates the performance of the macroscopic operation laws of the human brain.

At that time, the human experiments designed by the earthlings were mainly experiments on babies, because they could minimize distractions. Choose babies with similar cognitive development levels and let them recognize pictures for training, just like a one- or two-year-old child now, looking at a picture book to teach them what is a cat and what a dog is. This training is very similar to machine vision training for deep learning.

The sample size was then enlarged to give each baby a different ratio of right and wrong. The result is that babies with an error rate of close to % are the fastest to understand new things. For adults, the experiment cannot be designed for the time being, because there are too many distractions.

This leads to a surprising conclusion: what is the most effective way to learn?

The conclusion is: for living people, it is also a knowledge point, there are 15% of the content that you don't understand, and 85% of the basic knowledge that you understand.

At this time, your curiosity will be mobilized to the highest level, and your fear and rejection of the complete unknown will be suppressed to just the right level.

Why do so many scumbags learn from them? Why is the learning efficiency low? Is it not because of his grades, and there is no "right and wrong ratio between understanding and not understanding" that just fits the difficulty of the teacher's education?

Why are there so many paragraphs that say that the poor math student just picked up a pen in math class back then, and when he raised his head again, he no longer understood what the teacher was talking about?

Although this paragraph is a joke, many of the science scum are disconnected from the study area to the panic area bit by bit, and finally give up treatment.

At this time, if there is a tutor who knows that you have fallen into the panic zone, is willing to understand you, find out your level, give you a study zone that you are most comfortable with, and slightly adjust the difficulty, maybe these people's lives can be saved.

Many experienced gold medal teachers actually do this job, because the teachers actually understand the knowledge points in the book. The difference between a good teacher and a bad teacher is that a good teacher is experienced, and with a few questions, you can know what level the child is at now, how far behind he is, and what kind of difficulty and rhythm he should use to teach him according to his aptitude.

It's just that most good teachers only rely on experience, and have not systematically and accurately summarized data from a scientific point of view.

In fact, it is impossible for human learning to be accurate to % of the difficulty of not understanding.

But basically, a question or a piece of knowledge, packaged into a ratio of "seven questions, six right and one wrong", is already very efficient, at least can mobilize people's enthusiasm and curiosity for learning to more than 90% of the theoretical peak.

(The "right, six, and wrong" here is not a simple right-and-error question, but just to make a learning block of knowledge, six-sevenths of what you understand, and one-seventh that you don't understand.)

And this one-seventh should be inherited from the six-sevenths that I understood earlier. You can learn the last seven by summarizing and summarizing the first six-sevenths. Then draw the bread a little bit further, and then draw one-seventh of the ones you don't understand, and break them all. ）

Considering that the Chinese are the most utilitarian, they also attach the most importance to education and climb up.

Gu Yu believes that if there is an absolutely scientific method that allows people to master the flow and let themselves enter the state of forgetting about things and self at any time, then whether they love science, liberal arts students or science students, they will pay attention to this achievement.

Relying on the results of such a thesis, it is no problem to brush up on fame and brush up to the level of national godfathers. Because a layman can understand it and fully resonate.

"How should the experiment be designed?" asked Ma Yiyi, itchingly.

"You can work with those infant rehabilitation education providers to help adjust the right and wrong ratio of visual recognition training. Have children with similar scores on the pre-rehabilitation test do questions with different ratios of right and wrong to see who recovers faster

Rest assured, there is no moral pressure, because after we intervene, we will only allow these children to recover faster than they did if they had not intervened. The dataset you design will be more effective than the blind practice of not having a professionally designed dataset at present.

As for experimental resources, I'll help you use your connections to get in touch. When the accumulation of convolutional neural networks and deep learning algorithms is almost complete, your results can enter the node of joint detonation. ”

ps: There are already 3,000 words above, and the following routines don't spit out a few sentences.

I know that this chapter is not really closely related to the main story, but I chose to write it in detail.

The book is already like this, and I don't want to write it very coolly, anyway, it's all like these people who read it anyway.

I still have a simple idea, hoping that my book can raise the reader's IQ and do something for popular science. After reading my book, people who were not so good at learning are not only cool, but also more awesome in real life.

I don't want to harm young people, I hope that if there are students among my readers, they will be more successful in the next year than they are now, and they will still have the money to read my genuine books.

I don't want to make money on the demographic dividend.

Also, I think it's also a way to fight Xiao Baiwen. Because as long as there is one more person with education and improved IQ in the national reader audience, there will be one less person who looks at Xiaobai.

Therefore, improving the people's IQ and knowledge undoubtedly belongs to the category of fighting with Xiao Baiwen.

I'm not for you, I'm for myself.

I hope that readers who can even read popular science literature can benefit from it and live a more glamorous life in the future than readers who only read Xiaobaiwen.

Let's do long-term business.

Of course, no matter how loud the slogan is, I first have to admit that this book is really not well written, and I lack skills. So I didn't even bother to study because of my bad grades, a vicious circle.

But there is no way, it is impossible to learn a lot without knowing whether there is a future in a field.

Even if I write professionally and well-paced, I now look back and estimate that physics and cosmology are two or three thousand at a certain limit.